Message-ID: <428b865e0709180753t5f96c1eex2bed6d9d9a17c7fa@mail.gmail.com>
Date: Tue, 18 Sep 2007 09:53:28 -0500
From: "Evan Carroll" <lists@evancarroll.com>
To: "soni de" <soni.de@gmail.com>, pgsql-performance@postgresql.org
Subject: Re: Regarding COPY command from Postgres 8.2.0
In-Reply-To: <9f2e40a90709172131g26d681a1pf030910a1793f791@mail.gmail.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 7bit
Content-Disposition: inline
References: <9f2e40a90709140321l11e6cdb5g91d8f8bf0a43fd01@mail.gmail.com>
	<20070914120321.GA2615@mail.planetcobalt.net>
	<9f2e40a90709172131g26d681a1pf030910a1793f791@mail.gmail.com>

On 9/17/07, soni de <soni.de@gmail.com> wrote:
> We have upgraded postgres from 7.2.4 to 8.2.0.

This is one of the joys of 8.x over 7.2.x think of it like a different
sql product rather than an "upgrade." Its foundations are different.
7.4.x is still supported, and would have been a smoother upgrade for
you with less deprecations removed. To insert such a feature dependent
on column order without it being implied in the spec would be surely
be something deprecated quite quickly.

> We have program which executes COPY command and our new database is changed
> having some extra columns in some tables.
> Because of this, COPY commands are failing.
> So, we wanted the option to COPY the data without specifying column names.
>

The only way I can see a request, that allows for such bad-practice,
being approved would be if you were to use the '*' operator elegantly;
and, even then it would require you to update your code, only slightly
to a lesser degree than writing the column names.

Evan Carroll